OBJECTIVES: To describe the development of the Sheffield Preference-based Leg Ulcer patient questionnaire (SPVU-5D) and value the resulting health states.  METHODS: Patient and clinician interviews (n=31) were used to identify items for inclusion.  The psychometric performance of the items in the draft descriptive system were evaluated through a postal survey of venous ulcer patients (n=266).  Factor and Rasch analysis were used to guide the choice of the final items.  Preference weights for the final five item instrument were obtained through a time trade-off (TTO) valuation of a representative sample of the UK general population (n=160) using regression models incorporating individual level data.  RESULTS: A total of 16 items were identified for inclusion.  A response rate of 59% was obtained from the postal survey (n=156). Four items were excluded due to poor completion rates or floor/ ceiling effects.  Factor analysis using promax rotation showed a two factor solution with Cronbach's alpha coefficient of 0.91 for 0.92. Rasch analysis showed six disordered items, which had item levels merged and the model refitted.  The final descriptive system consisted of five items with between three and five levels.  A fractional factorial design was employed to select 25 health states from the 720 defined by the final instrument.  Predicted health state values were calculated and compared to the actual values derived from the TTO valuation. The best performing model (OLS model) predicted 25 out of the 26 health states valued in the TT0 to within 0.1, and the predicted value within 0.03, 69% of the time and within 0.05 92% of the time. The mean absolute error was 0.02. CONCLUSIONS: A bottom-up approach was used to design SPVU-5D.  It is the only existing PRO which has utility values assigned to the health states it describes and so can be used within economic analysis of interventions for venous ulceration.